Test-Retest Apparent Diffusion Coefficient Reproducibility in Head and Neck Cancer Using a 1.5-T MR-Linac

Area of Science:

• Radiology
• Oncology
• Medical Physics

Background:

• Diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) measurements are crucial for assessing head and neck squamous cell carcinoma (HNSCC) response to radiation therapy.
• MRI-guided linear accelerator (MR-linac) systems offer integrated imaging and treatment capabilities, but their reproducibility in quantitative MRI metrics like ADC needs evaluation.

Purpose of the Study:

• To evaluate the reproducibility of ADC measurements in HNSCC using a 1.5-T MR-linac system.
• To determine if lesion volume or time between scans affects ADC measurement reproducibility.

Main Methods:

• Retrospective analysis of 37 HNSCC patients undergoing two echo-planar imaging DWI scans on a 1.5-T MR-linac before radiation therapy.
• Calculation of mean and median ADC values and volumes, along with absolute and percent reproducibility coefficients (RCs).
• Linear regression and F tests were used to assess the impact of lesion volume and scan interval on reproducibility.

Main Results:

• The study included 34 primary tumors and 55 lymph nodes. Median ADC values for tumors and lymph nodes were comparable (approximately 1.16-1.19 × 10^-3 mm²/sec).
• Reproducibility coefficients (RCs) for mean ADC were 0.365 × 10^-3 mm²/sec for tumors and 0.355 × 10^-3 mm²/sec for lymph nodes (percent RCs of 29.9% and 31.1%, respectively).
• No significant correlation was found between lesion volume or scan interval and ADC reproducibility, though a trend towards poorer reproducibility in smaller lesions was observed.

Conclusions:

• The 1.5-T MR-linac DWI sequence demonstrates acceptable reproducibility for detecting substantial ADC changes in HNSCC.
• While generally reproducible, the system may miss some clinically significant, smaller ADC variations, particularly in smaller lesions.
• Further optimization may be needed to enhance sensitivity for detecting subtle treatment response changes in HNSCC using MR-linac technology.